The purpose of this project is to utilize genomic data, primatily those concerned with sequences of complete intergenic regions og genomes, in order to learn more about functions of non-coding eukaryotic DNA. For this purpose, we developed a software tool OWEN for rapid, hierarchical alignment of very long sequences such that similarity between them can be represented by a chain of successive local alignments. Alignments obtained in this way were used for the following studies: - quantitative description of selective constraint revealed by Homo - Mus and Caenorhabditis elegans - C. briggsae alignments, - analysis of domestication of ancient transposons MIR and L2 in mammals, - classification of highly conserved, common fragments in mammalian non-coding DNA,